Degradation of skin elastin in ageing and disease - an ex vivo quantitative proteomic approach

(1)

Degradation of skin elastin in ageing and disease:

An ex vivo quantitative proteomic approach

Dissertation

zur Erlangung des

Doktorgrades der Naturwissenschaften (Dr. rer. nat.)

der

Naturwissenschaftlichen Fakultät I – Biowissenschaften –

der Martin-Luther-Universität

Halle-Wittenberg,

vorgelegt

von Frau Angela Cristina Mora Huertas

geb. am 04. März 1980 in Guateque, Kolumbien

Gutachter:

1. Prof. Dr. Dr. Reinhard. Neubert 2. Prof. Dr. Dierk Scheel

3. Assoc. Prof. Dr. .Laurent Duca

(2)

(3)

Acknowledgments

I

Acknowledgments

Thank God for blessing each moment in my life and most of all for allowing me to achieve this goal.

First of all, I would like to show my deepest gratitude to Dr. Andrea Heinz. Her timely guidance, constructive criticism and scientific approach have helped me in a great extent to finish this work. Thanking you is not enough, but it is the best way to pay tribute to a person whose endless support, kind and understanding spirit played a key role throughout this process.

I am also very grateful to Dr. Christian Schmelzer for accepting me as a member of his research group and introducing me to the fascinating topic of the elastin and the mass spectrometry. It has been delightful to broaden my knowledge around your kindness, expertise and help.

Special thanks to Prof. Dr. Reinhard Neubert for the opportunity of doing my doctoral project in the AG Biopharmazie and for his permanent assistance with the administrative issues.

On the part of Leibniz Institute for Plant Biochemistry of Halle; I would like to thank Dr. Wolfgang Hoehenwarter for providing the ‘Progenesis Q.I’ software as well as Miss Petra Majovsky for her measurements with the Orbitrap Q Exactive Mass Spectrometer. I would like to acknowledge the Martin Luther University Halle-Wittenberg team, specially Dr. Frank Heyroth for his thoughtful help with scanning electron microscopy on the Interdisciplinary Center for Materials Science and Prof. Dr. Johannes Wohlrab from the Department of Dermatology and Venereology for the supplying the skin biopsies samples.

My thanks and appreciations also to my colleagues Jing, Christoph and Tobias, who have willingly helped me out in innumerable ways; to Angelica Avila for her hard work and enthusiasm during our lab time; and to Frau Manuela Woigk for her assistance in daily lab work.

(4)

Acknowledgments

II Thanks to the Universidad Nacional de Colombia in Bogota for giving me the opportunity to foster my academic development and to the Department of Pharmacy’s professors for your kind support.

I would like to express my deepest gratitude to my mentors, Prof. Noralba Sierra, Roberto Pinzon and Luis F. Ospina for their wisdom, knowledge, trust and care.

Thanks to Colombian Institutions: Departamento Administrativo de Ciencia, Tecnología e Innovación (Colciencias) and Fundación para el Futuro de Colombia (Colfuturo) for the funding sources and its management.

On my family’s side, my deepest gratitude to my parents, Elisenia and Rafael; your teaching, love, and encouragement made it possible. To my siblings Neyla, Claudia, Ricardo and particularly to Rafael and his family, Luisa, Catalina and Sergio, for all the love and support received. Also to Karen and Tulia for your kind encouragement, as well as my uncles, aunts and cousins.

My specials thanks to Martin for his patience and support during this writing time, and for encouraging me to strive towards my goal. Thanks for all the special moments that we have shared.

My extended hug and warm thanks to my friends here in Germany, specially to Alejandra, Hina, Marisol, Yenny, Efrem, Elkin, Julio, Mauricio and Walter for their support, care, advice and for making my stay in Germany homey and comfortable. Special thanks to my friends in Colombia, particularly to Carolina, Martha, Sandra and Yoshie; your encouragement, enthusiasm and friendship made this journey easier.

Finally, I want to express my deepest gratitude to my Uncle Tito for showing me that it is possible to achieve this dream and for lightening my way when the dark times arrived.

(5)

Table of Contents

III

4.4.3 Differences between elastin peptides released from elastin isolated from skin of WBS patients and healthy individuals ... 77 4.4.4 Release of potentially bioactive peptides from elastin isolated from WBS patients and healthy donors. ... 78 4.4.5 Classification of samples according to the elastin changes in WBS patients and healthy individuals ... 79

5 Discussion ... 82

5.1 Workflow suitability to distinguish changes in abundance of elastin hydrophobic

peptides. ... 82 5.2 Susceptibility of human skin elastin towards degradation by biologically relevant

proteases ... 84 5.2.1 Elastase activity of NEP and its relation with previous elastic fibres damage ... 85 5.2.2 Degradation of TE and skin elastin samples by CG and MMP-9 ... 87 5.2.3 Enzymatic susceptibility of the elastin domains analysed through their

degradation by CG and MMP-9 ... 89 5.2.4 Age-related differences in the elastin susceptibility towards enzymatic

degradation ... 92 5.2.5 Peptides with bioactive sequences released from different types of skin elastin samples 95

5.3 Structural changes of human elastin during skin ageing ... 97 5.3.1 Elastic fibres morphology and its susceptibility towards enzymatic degradation . 97 5.3.2 Characterisation of elastin peptides released from elastin obtained from

differentiated aged individuals ... 98 5.3.3 Release of potentially bioactive peptides from elastin during skin ageing... 100 5.3.4 Classification of samples according to the sources of elastin degradation ... 101 5.4 Molecular changes of human skin elastin from patients with Williams-Beuren Syndrome and healthy individuals ... 103

5.4.1 Elastin content of skin and elastic fibre morphology ... 103 5.4.2 Elastin susceptibility towards enzymatic cleavage... 103 5.4.3 Differences between elastin peptides released from elastin isolated from skin of WBS patients and healthy individuals ... 104 5.4.4 Release of potentially bioactive peptides from elastin isolated from WBS patients and healthy donors. ... 105 5.4.5 Classification of samples according to the elastin changes in WBS patients and healthy individuals ... 106

6 Conclusions ...108 Appendix ... XVII

(8)

Table of Contents

VI

Appendix 1. Workflow suitability to distinguish changes in abundance of elastin hydrophobic peptides ...XVII Appendix 2. Susceptibility of human skin elastin towards degradation by biologically relevant proteases ... XVIII Appendix 3. Structural changes of human elastin during skin ageing ... XXVII Appendix 4. Molecular changes of human skin elastin from patients with Williams-Beuren Syndrome and healthy individuals ... XXXII

Bibliography ... XVII Curriculum vitae ... XXXVI List of publications ... XXXVII Declaration ... XXXIX

(9)

List of Acronyms

VII

List of Acronyms

a.u. Arbitrary units ACN Acetonitrile

ACP Allysine aldol condensation product ADCL Autosomal dominant cutis laxa APMA p-aminophenylmercuric acetate AUC Area under the curve

CE Children skin elastin CG Cathepsin G

CHCA alpha-cyano-4-hydroxycinnamic acid CHO Chinese hamster ovary cells

CID Collision-induced dissociation CV Coefficent of variation

DDA Data-dependent acquisition ddH2O Double distilled water

DEJ Dermal-epidermal junction dLNL Dehydrolysinonorleucine DMSO Dimethyl sulfoxide

EBP 67-kDa Elastin-binding protein ECM Extracellular matrix

EDPs Elastin-derived peptides ESI Electrospray ionization FA Formic acid

FC Fold change

FDR False discovery rate

HCA Hierarchical cluster analysis

HPCA Hierarchical principal component analysis HPLC High performance liquid chromatography HUVEC Human umbilical vein endothelial cells IT Ion trap

(10)

List of Acronyms

VIII LID Laser-induced dissociation

LOX Lysine-6-oxidase

MALDI Matrix-assisted laser desorption/ionization MMP Matrix metalloproteinase

MS Mass spectrometry NEP Neprilysin

OE Old adult elastin PC Principal component

PCA Principal componant analysis PE Pancreatic elastase

PMNL Polymorphonuclear leukocytes Q Quadrupole

SA Sinapic acid

SEM Scanning electron microscopy SFE Skin fibroblast-derived elastase SVAS Supravalvular aortic stenosis TE Tropoelastin

TFA 2,2,2-trifluoroacetic acid TOF Time-of-flight

Tris 2-Amino-2-hydroxymethyl-propane-1,3-diol UV Ultraviolet

w.a.m. Weighted arithmetic mean WBS Williams-Beuren Syndrome

(11)

List of Figures

IX

List of Figures

Figure 1. Skin and its cellular and extracellular matrix components. ... 1

Figure 2 Elastin fibres ... 4

Figure 3. Elastic fibre assembly. ... 5

Figure 4. Schematic diagram of the domain structure of human tropoelastin. ... 7

Figure 5. Model of the nanostructure of full-length and assembly of Tropoelastin. ... 10

Figure 6. Phenotype of intrinsic ageing versus photoaging. ... 15

Figure 7. Principal components of a mass spectrometer. ... 18

Figure 8. Label-free quantification (LFQ) methods. ... 24

Figure 9. Normalised abundance of selected elastin peptides depending on the elastin concentration ... 40

Figure 10. PCA scores plot of samples with different elastin concentration ... 41

Figure 11. Variable graph of peptides quantified in samples with different elastin concentration. ... 41

Figure 12. MALDI-TOF mass spectra of TE and skin elastin samples degraded by NEP. . 43

Figure 13. Cleavage sites identified after digestion of TE and skin elastin samples with NEP. ... 44

Figure 14. Normalised number of cleavage sites identified in samples of TE and skin elastin digested with NEP. ... 45

Figure 15. MALDI-TOF mass spectra of TE and skin elastin samples degraded by CG and MMP-9. ... 47

Figure 16. Cleavage sites identified after digestion of TE and skin elastin samples with CG (A) and MMP-9 (B). ... 49

Figure 17. Sequence coverage obtained from peptides quantified after 6 h, 12 h and 48 h in CG digests. ... 50

Figure 18. Changes in normalised abundance of selected TE peptides depending on sampling point. ... 51

Figure 19. Sequence coverage obtained from peptides quantified after 6 h, 12 h and 48 h in MMP-9 digests. ... 52

Figure 20. Normalised number of cleavage sites identified in samples of skin elastin digested with CG and MMP-9, respectively. ... 54

(12)

List of Figures

X Figure 21. Sum of normalised abundance by domain of elastin peptides solubilised from skin elastin digested with CG. ... 55 Figure 22. Sum of normalised abundance by domain of elastin peptides solubilised from skin elastin digested with MMP-9. ... 56 Figure 23. Degradation of different domains of elastin by CG and MMP-9, determined through a linear model. ... 58 Figure 24. Profile of quantifiable peptides obtained after digestion of TE and skin elastin samples with CG and MMP-9. ... 60 Figure 25. Amount of elastin peptides solubilised after 48 h, and quantified using an UV spectrophotometric method. ... 61 Figure 26. Modelling of elastin peptides released from mature elastin from domain 28.62 Figure 27. Scanning electron micrographs of human skin elastin obtained from

differential aged healthy individuals ... 66 Figure 28. Elastin peptides identified from PE digests of skin elastin samples from

differential aged individuals. ... 67 Figure 29. Changes in normalised abundances of selected elastin peptides depending on the age of the donor. ... 68 Figure 30. First cleavage points in skin elastin isolated from differential aged individuals. ... 70 Figure 31. PCA scores plot of skin elastin samples isolated from differential aged

individuals. ... 71 Figure 32 Variables graph of 55 elastin peptides with age-related changes, based on PC1 and PC2. ... 72 Figure 33. Dendrogram of skin elastin samples isolated from differential aged

individuals. ... 73 Figure 34. Morphological characterisation of skin elastin isolated from healthy donors and WBS patients. ... 75 Figure 35. Cleavage sites identified after PE digestion of skin elastin samples from WBS patients and healthy donors. ... 76 Figure 36. PCA scores plot of skin elastin samples isolated from WBS patients and

healthy individuals... 79 Figure 37 Variable graph of 64 elastin peptides with significant differences between WBS patients and healthy individuals based on PC1 and PC2... 80

(13)

List of Figures

XI Figure 38. Dendrogram of skin samples isolated from WBS patients and healthy

individuals. ... 81

Figures in the appendix

Figure A-1. Normalised abundance of peptides quantified after 6 h, 12 h and 48 h in TE samples digested with CG... XVIII Figure A-2. Normalised abundance of peptides quantified after 6 h, 12 h and 48 h in CE samples digested with CG... XIX Figure A-3. Normalised abundance of peptides quantified after 6 h, 12 h and 48 h in OE samples digested with CG... XX Figure A-4. Normalised abundance of peptides quantified after 6 h, 12 h and 48 h in TE samples digested with MMP-9. ... XXI Figure A-5. Normalised abundance of peptides quantified after 6 h, 12 h and 48 h in CE samples digested with MMP-9. ... XXII Figure A-6. Normalised abundance of peptides quantified after 6 h, 12 h and 48 h in OE samples digested with MMP-9. ... XXIII Figure A-7. Degradation of different domains of skin elastin determined through a linear model according to each enzyme. ... XXIV Figure A-8. Sum of normalised amount of peptides obtained after digestion of TE and skin elastin samples with CG and MMP-9, according to biological replicate. ... XXV Figure A-9. Sum of normalised amount of peptides obtained after digestion of TE and skin elastin samples with CG and MMP-9, according to instrumental replicate. ... XXV Figure A-10. Changes in normalised abundance of elastin peptides depending on the age of the donor (part A). ... XXVIII Figure A-11. Changes in normalised abundance of elastin peptides depending on the age of the donor (part B). ... XXIX

(14)

List of Tables

XII

List of Tables

Table 1. In vitro biological activities reported for some EDPs. ... 14

Table 2. Chemicals ... 28

Table 3. Protein and Enzyms ... 28

Table 4. Buffer and reagent composition ... 29

Table 5. Instruments ... 30

Table 6. Occurrence of different amino acids at the substrate positions P1-P4 and P1´-P4´ after digestion of TE and skin elastin samples with NEP. ... 46

Table in the appendix

Table A-1. Elastin peptides normalised abundance among different elastin

concentrations. ... XVII Table A-2. Most abundant peptides containing bioactive sequences that were identified and quantified after digestion of human skin elastin by CG and MMP-9. ... XXVI Table A-3 Skin samples from differential aged healthy individuals. ... XXVII Table A-4. Elastin peptides that showed significant differences in their normalised abundances among different elastin samples. ... XXXI Table A-5. Skin samples from WBS patients and healthy individuals analysed by

nanoHPLC-nanoESI-QqTOF MS and LFQ. ... XXXII Table A-6 Elastin peptides that show significant differences in their normalised

abundances between skin elastin samples isolated from WBS patients and healthy individuals. ... XXXIV

(15)

Abstract

XIII

Abstract

The degradation of elastin, the most abundant protein in the elastic fibres, has a pivotal role in the loss of the integrity and functionality of the skin during the ageing and some diseases such as Williams-Beuren Syndrome (WBS). Previous studies of elastin degradation have reported, which elastases are involved in this process and how they degrade elastin. The aim of this study was to investigate the degradation of skin elastin at the molecular level by evaluating the enzymatic susceptibility and age-related changes of its hydrophobic domains, and the effect of the skin ageing and WBS on their morphology and susceptibility towards enzymatic degradation.

Elastin fibres were isolated from skin biopsies of differently aged healthy individuals and WBS patients. Their morphology was characterised by scanning electron microscopy. Mass spectrometric techniques were used to investigate the molecular-level structure of elastin. For the first time, label-free quantification (LFQ) and statistical analysis were applied to study quantitative changes in the peptides released during the elastin degradation.

The analysis of the cathepsin G (CG) and matrix metalloproteinase 9 (MMP-9) digests revealed that thirteen domains in the N- and C-terminal regions of tropoelastin (TE) are particularly susceptible to enzymatic degradation. The degree of degradation of each domain was the result of the interaction of the type of enzyme, the integrity of the elastin fibre and incubation time.

The age-related loss of integrity of the fibrillar elastin favoured the elastolytic activity of neprilysin (potential skin fibroblast-derived elastase), while it decreased the efficacy of MMP-9 degradation. Elastase capacity of CG is not influenced by these changes.

Insights in the ageing process of the elastin fibres were obtained through marker peptides which showed an age-related increase or decrease in their abundances. Domains 18, 20, 24 and 26 were strongly cleaved with increasing age. The susceptibility towards enzymatic cleavage of the N-terminal and central regions of the TE was increased by the extrinsic ageing. It also stimulates the strong decomposition of elastin fibres observed in sun-exposed skin samples.

(16)

Abstract

XIV Skin elastin fibres obtained from WBS patients differed from the ones obtained from healthy individuals in three features: firstly, their high susceptibility to enzymatic cleavage, secondly, their lower proline hydroxylation degree and thirdly, the diminished total amount in the tissue. Moreover, the accelerated damage of these elastin fibres resembled the effect of extrinsic skin ageing in old adult donors.

(17)

Theoretical Background

1

1 Theoretical Background

1.1 Human skin and extracellular matrix

The skin is the largest organ of the human body and covers 1.6 m2_{of surface area}

[10]. It helps to maintain four essential body functions such as sensation, retention of moisture and prevention of permeation or other molecules loss, regulation of body temperature and protection of the body from external factors [11]. Additionally, the skin is a dynamic structure which involves multi-directional stretch and compression, allowing for low friction gliding movement [12].

Morphologically, human skin is composed of three distinct layers: epidermis, dermis and hypodermis [11, 13] (Figure 1). The epidermis is avascular and mainly constituted by epidermal keratinocytes and their mature cells, the cornified cells [11]. The structure beneath the epidermis, the dermis, is vascularized and relatively acellular. These two layers are united through the dermal–epidermal junction (DEJ) [4]. The last skin layer is the hypodermis, which mainly consists of loose connective tissue and it is particularly rich in proteoglycans and glycosaminoglycans [12].

Figure 1. Skin and its cellular and extracellular matrix components.

(A). In mammalian skin, the dermis supports the epidermis. The dermis is structurally composed of the papillary and reticular layers. The deeper layer, hypodermis, is not shown in the figure. (B). Cellular components of the epidermis (keratinocytes) and dermis (fibroblasts). (C). ECM constituents. The fibre-forming structural components (collagens and elastic fibres); the non-fibre-fibre-forming molecules (proteoglycans and glycosaminoglycans); and different matricellular proteins. The figure was modified from Naylor et al. [4].

(18)

2 There are two regionally distinct areas in the dermis. The papillary layer has small diameter collagen fibres interspersed with thin elastic fibres [12], and they are richly supplied with capillaries, sensory nerve endings and cytoplasm [11]. The reticular layer, which is predominantly made up of collagen fibres, is less densely packed and organised into large interwoven fibre bundles of branching elastic fibres; these fibres, in turn, form a superstructure around the collagen fibres [12]. The majority of cells in the human dermis are fibroblasts that are thought to be responsible for synthesising the dermal extracellular matrix (ECM) proteins [4].

As in other tissues, ECM components of the skin maintain its structural integrity and participate actively in numerous aspects of cellular regulation [14]. ECM components could be sorted into three categories (Figure 1): (i) the fibre-forming structural components are made of collagens and elastic fibres. Collagens confer tensile strength to the tissue [15]. Collagens type I and III are widely distributed inside papillary and deep reticular dermis, while collagen type VII is restricted to the DEJ [4]. The elastic fibre system is fundamental in mediating tissue resilience and elasticity [15-17], and it forms a three-dimensional meshwork that spans from the papillary down to the deep dermis [18]. In the reticular dermis there are three different types of elastic fibres: oxytalan, elaunin and elastic fibres. Oxytalan fibres, which do not contain amorphous elastin, form a fine branch such as fibrillin-rich microfibrils. The elaunin fibres are arciform microfibrils with a poor elastin core. Finally, the elastic fibres are thick fibrillin-rich microfibrils with an elastin-rich core [12, 19, 20]. (ii) The non-fibre-forming molecules include proteoglycans and glycosaminoglycans, such as hyaluronic acid and chondroitin sulphate glycosaminoglycan. The function consists in hydrating the skin due to their capacity to create a charged, dynamic and osmotically active space [4, 15]. (iii) The matricellular proteins such as osteopontin, secreted protein acidic and rich in cysteine (also known as osteonectin), tenascin-C, fibulins, and the CCN family, do not have a structural function but interact in autocrine or paracrine cell-matrix signalling [15]. Tracy et al. reviewed the skin ECM components and their interactions [15].

(19)

3

1.2 Elastic fibres

1.2.1 Structure of elastic fibres

Elastic fibres are the major insoluble ECM structures that endow connective tissues subjected to repetitive distension and physical stress with resilience, permitting their low-strain mechanical response [17, 21]. These fibres are found in several tissues such as skin, lungs, alveoli, arteries, veins, the urinary tract, eye (both the cornea and suspensory ligament), fibrillar and articular cartilage, and specialised tendons such as the ligamentum nuchae [11, 13].

The assemblage of elastic fibres comprises various distinctive microfibrillar structural glycoproteins, while elastin forms an inner amorphous core of the compound elastic fibres, and one enzyme [19, 22]. Elastin-associated microfibrils include fibrillin 1-3 [20] and some associated molecules as the latent TGF-β binding proteins (LTBPs), ADAMTS (a disintegrin and metalloproteinase with thrombospondin motifs), MAGPs (microfibril-associated glycoproteins), fibulin 2–5 and the lysyl oxidase enzyme (LOX) [17, 20, 23-26]. LOX is involved in the elastin cross-linking [27] while the elastin-associated microfibrils and fibrillin-microfibrils contribute to elastic fibre assembly and function [20], however, the last one does not play a major role in the elasticity activity [28]. For a review of microfibrillar structure, see [17, 29, 30].

Elastin is the most abundant protein in the elastic fibres; it constitutes approximately 90 % of the mature structure of them [31]. As it was mentioned previously, it provides extensibility/elastic recoil [32], and resilience [15-17] to the different tissues and it is, thus, critical for their long-term function [9, 33-35]. This protein has an entropy-based elasticity mechanism, in which the hydrophobic hydration and the release of hydration waters play the dominant role [36-39]. The dynamic nature of elastin´s hydrophobic domains contributes in less extent to the elasticity [37, 40, 41].

The amount of elastin and general architecture of the mature elastic fibres have tissue-specific building arrangements that reflect different elastic requirements [14, 16, 42]. For instance, the skin contains between 2 % and 5 % of elastin, on the contrary, the aorta´s composition is between 30 % and 57 % of the protein (percentages based on dry weight of the tissue) [16, 31]. Furthermore, dermal elasticity relies on integrated

(20)

4 networks of thick reticular elastic fibres whereas the arterial elastic fibres form concentric lamellar layers that support vascular elastic recoil [14, 20]. Although the histologic structure of elastic fibres differs among tissues, the fine detail of insoluble elastin is apparently similar regardless of location [14]. Macroscopically, elastin exhibits an amorphous appearance. At supramolecular level, the protein has a fibrillar substructure comprised of parallel-aligned ≈ 5 nm thick filaments, isolated or laterally aligned. They appear to be hierarchically organised as twisted-rope fibres and fibrils [18, 43, 44] (see Figure 2). For a review of the supramolecular organisation of elastin and other elastin-related compounds, see Pepe et al. [44].

1.2.2 Elastogenesis

In humans, the elastic fibres are formed around the time of mid-gestation and the maximum production is reached near birth and during the early neonatal period [14]. The elastin deposition is nearly complete in the first decade of life [45]. In mature organs and tissues, elastin synthesis is repressed by post-transcriptional factors [46]. The low turnover and lack of continued elastin production upon maturity reflect the extreme durability and long half-life of elastic fibres that reach the human lifespan of around 74 years [47]. However, tropoelastin (TE) expression may be reinitiated in response to wounding [47, 48] or exposure to ultraviolet (UV) radiation [49, 50]; elastin production is aberrant in these cases and does not lead to the formation of normal elastic fibres.

Figure 2 Elastin fibres

The scanning electron micrographs of elastin fibres isolated from the human skin of a volunteer at the age of 6. Figure adopted from Schmelzer et al. [149].

(21)

5 The in vivo formation of elastic fibres can be described in two independent steps: the formation of microfibrils and the arrangement of elastin core [51]. Figure 3 shows all the steps involved in the elastogenesis. A detailed description of elastogenesis can be found in Schmelzer et al. [3].

During elastic fibre synthesis, the microfibrils, mainly consisting of fibrillin-1, appear before the amorphous core and are believed to act as a scaffold for the deposition of elastin [13, 24, 52]. The TE is encoded by a gene on chromosome 7q11.2 in humans. Depending on anatomical location, TE (70 kDa) is secreted by different elastogenic cells such as fibroblasts, smooth muscle cells, chondroblasts, mesothelial cell, endothelial cells, and auricular chondrocytes [3, 14, 20, 45]. After that, TE binds to a 67-kDa elastin-binding protein (EBP) to avoid its intracellular self-aggregation and degradation, and it is transported to the membrane where it is released into the extracellular space [53, 54].

Figure 3. Elastic fibre assembly.

(1) Microfibrillar array in extracellular space. (2) Synthesis and binding of TE to the EBP in the rough endoplasmic reticulum (ER) (3) Transport of EBP-TE complex through the Golgi apparatus and secretion to the cell membrane. (4) The release of TE from EBP and formation of globules at the cell surface through its cross-linking mediated by lysyl oxidase or lysyl oxidase-like enzymes (5) Deposition of TE clusters onto the microfibrillar array. (6) Fusion of elastin aggregates into larger assemblies and further cross-linked to eventually form the elastic fibre. Figure and text adopted from Schmelzer [3].

(22)

6 Small cell surface-associated TE globules appear and increase in size with time (micro assembly) [52] through an inverse temperature transition termed coacervation. In it, the molecules are concentrated and aligned as result of multiple and specific interactions of individual hydrophobic domains of TE [55, 56]. Details of coacervation process are reviewed in Yeo et al. [56].

After the elastin globules reach a critical size, they are eventually transferred to pre-existing microfibrillar fibres in the ECM where they coalesce into larger structures (macro assembly) [52]. The TE molecules are covalently bound to each other through the cross-linking process [57-59]. It is initiated by the oxidative deamination of lysine side chains by the enzyme protein lysine-6-oxidase (LOX; EC 1.4.3.13) to obtain the α-aminoadipic δ-semialdehyde (allysine) [60, 61]. Then, the allysine spontaneously condensates with another allysine to form allysine aldol condensation product (ACP). Another possible reaction is between allysine and the amine of an unmodified lysine side chain through a Schiff base reactions to form dehydrolysinonorleucine (dLNL). ACP and dLNL can then spontaneously condense with each other, or with other intermediates to form desmosine or its isomer, isodesmosine [58, 61, 62]. The increase in complexity of this inter- and intra-chain cross-links is thought to progress as the fibre matures and ages [63].

1.2.3 Tropoelastin and elastin structure

The human TE gene possesses 34 exons, but almost 30 % of them undergo alternate splicing [64] in a cassette-like fashion, in which an exon is either included or deleted, but rarely divided [65]. Splicing could be tissue-specific and developmentally regulated [51, 66]. It has been suggested that among tissues, the alternate splicing may influence the elastic fibre resilience, the fibre assembly [14, 64, 67] and the interaction of TE with other matrix molecules and cells [68]. As result of alternative splicing, TE can lose domains 22, 23, 24, 26A, 30, 32 and 33 [51, 69, 70]. However, in human skin elastin the absence of exon 26A was confirmed [71].

An additional intracellular post-translational modification that TE undergoes is the cleavage of the signal peptide in the endoplasmic reticulum [72], which apparently contributes to the correct folding of TE and its coacervation [73]; this process has to occurred before placement of the monomer at the ECM [72]. Another modification is the

(23)

7 hydroxylation (HyP) of some proline residues (about 1% HyP in elastin) in the endoplasmic reticulum [45, 74]. The role of hydroxylation is not completely understood [16, 75]. It seems to be related to different functions such as resistance to mechanical stresses of the organs and elastin assembly; it could also improve elastin’s resistance towards proteolytic degradation [75].

Domain structure of the protein (see Figure 4) is a reflection of the exon organisation of the gene [61]. TE arrangement is made up of alternating hydrophobic regions and hydrophilic cross-link domains. Domains 1 and 36 are the signal sequence and the C-terminal domain respectively [76]. Although the amino acid compositions of all elastins have similar general characteristics, there is a variation in composition between tissues and between species [42, 69]; mammalian and avian elastins lack the amino acids histidine and methionine [61, 77].

In mammalian elastin, the hydrophobic domains are characteristically rich in glycine (≈ 33 %), alanine (≈ 24 %), valine (≈ 13 %) and proline (≈ 10 %). They have a noticeable variation in the length and composition among the mammalian elastins [61, 77]. Hydrophobic domains include many short tandem repeats and quasi-repeat sequences [78], such as PGVGVA [64, 79], which are formed by the pairwise combination of the fragments PGV, GVA, GV and GGV [80]. These domains could be classified as glycine-rich, localised mainly in the C- and N-terminal regions and proline-rich domains are found mostly in the central part of the molecule (see Figure 4). The last one is usually referred

Figure 4. Schematic diagram of the domain structure of human tropoelastin. The figure was adapted from Tamburro et al. (2003) [1].

(24)

8 as proline-rich domains because of their higher number of Pro residues, even if a considerable number of Gly residues are present [1]. Hydrophobic sequences are associated with the entropic mechanism of elasticity [36-41]. Moreover, they contain some interaction sites that are necessary for the aggregation of TE [56]. Furthermore, peptides derived from hydrophobic domains can interact with cell receptors, induce cell signalling and be involved in several pathologies [81].

In contrast, the cross-linking domains contain mostly polar sequences containing Ala and Lys in the form of -AlaAlaLys- and -AlaAlaAlaLys-, which are the cross-linking sequences of elastin [64, 82]. The length of cross-linking segments is highly conserved among mammalian elastins, indicative of a strong functional requirement [61]. These domains are classified into two groups: KA domains, which are rich in Ala and located in the central and C-terminal region, and the KP domains that contain Pro and are confined principally to the first third of the molecule [40]. The cross-linking process is extremely efficient. Most of the Lys residues are involved into cross-links (only ≈ 5 of the ≈ 40 Lys residues do not participate in some form of cross-link), and there are very few charged residues in elastin [65, 83]. The cross-linking of these domains confers fundamental mechanical properties to the elastin such as resistance to rupture, reversible deformation and high resilience [40].

Structural analysis of the elastin is complex due to its unique properties such as hydrophobicity, insolubility in common solvents [32, 40, 61, 79, 83], the high mobility of the elastin backbone [32, 37, 65, 84], and its lack of crystallisation [85]. Little is known about how the protein is assembled at the molecular level [37, 40, 83, 84]. The structural analysis has been focused on investigating elastin solubilized by oxalic acid (α-elastin) [86-89], potassium hydroxide (κ-elastin) [84, 90], TE [32, 55], or elastin-like peptides [41, 73, 91]. Studying structural motifs is based on fractal properties of elastin, which indicate that the property of statistical self-similarity characterises the protein. Accordingly, short sequences show molecular and supramolecular features very similar to those of the whole protein [73, 91, 92]. The different analytical approaches used are circular dichroism [41, 73, 86, 87], Fourier transform infrared [41, 93, 94], fluorescence and nuclear magnetic resonance (NMR) spectroscopies [41, 91, 94], Raman spectroscopy [93] as well as X-ray diffraction methods [41] or molecular dynamics simulations [36, 37, 91].

(25)

9 As results of the different structural analysis previously mentioned, it has been suggested that elastin’s sequence has a heterogeneity in its secondary structure [95]. Hydrophobic elastin-derived polypeptides show a propensity to adopt β- (type I, II and VIII) and γ-turn conformations [36, 37, 40, 73, 93, 96-98], and β-strand [36, 73, 84, 93], PPII [40, 41, 73, 91, 99] and unordered structures [40, 41, 84, 88, 89]. KP domains adopted similar conformations [40, 100], and the KA domains mostly show α-helix conformations due to the cross-link requirement [73, 84, 89, 93, 100]. Analysis of full-length monomeric TE suggests that it contains α-helix (3 % - 10 %), β-structures including β-strand and turns (45 % - 60 %), random coil secondary structures (> 40 %) [88, 90, 101] and the presence of polyproline II structure [1, 91, 102]. For a review of the structure of elastin and elastin-like polypeptides, from features of domain organisation to the supramolecular organisation of the fibre and structural flexibility see [95].

Regarding the tertiary structure adopted by TE in solution, there is not consensus either [32, 95, 103]. It has been suggested that the monomer is flexible, really dynamic [32, 95] and it has a high disordered backbone [32, 104, 105]; it looks like a transient structural form [32, 95, 103]. The structure of TE in solution has also been described as a ´thermodynamically unfolded premolten globule` state [95, 103], in which the monomer contains pockets of hydrophobic clusters, which are solvent accessible and not confined to a molecular core [32].

Taking into account the results of small-angle X-ray and neutron scattering experiments, Baldock et al. suggested a model for the nanostructure of full-length TE, which describes it as an asymmetric molecule with clearly distinct regions [9] (see Figure 5A). The same authors also propounded a head-to-tail model for the assembly of the TE. In it, the alignment and n-mer propagation of TE happen through the interaction of proximal domains 19 and 25, which are donated by one TE, and the domain 10 from a second molecule (Figure 5B). This model is based on the three cross-link domains (10, 19, and 25) and the junction with the longest molecular springs (domains 18, 20, 24 and 26) [9]. Chains of the three cross-link domains were also identified to be joined by one desmosine and two lysinonorleucine cross-links in porcine elastin [58].

(26)

10 There is no consensus about the secondary and tertiary structure of elastin either [40, 106]. Elastin-like peptides only correspond to small parts of the elastin, and their relevance to the protein structure is doubtful [3, 107, 108]. In mature elastin, hydrophobic clusters remain partially solvent accessible, to a lesser extent, though [109]. It has been hypothesised that elastin has a limited formation (if any) of a stable and buried core due to the solvent exposition and mobility of its hydrophobic domains [36, 37, 95]. It also has been hypothesised that elastin has disordered structures [39].

1.2.4 Structural role of some tropoelastin domains

As result of the analysis of elastin-like peptides, some special roles of different TE domains have been suggested. Overall, the N-terminal elastic coil region has been associated with the elastic fibre assembly [110] and elasticity due to its spring-like properties [9]. Sequences from exon 2 to 7 [73] and 16 to 18 may also contribute to TE self-association [111-114]. Specific intra-molecular cross-links were confined to the region encoded by exons 6-15 [51]. Notably, domain 12 participates in cross-linking with domain 23 and could play a major role in the formation of mature elastin; domains 13 and 14 are also more exposed and available for reaction predominantly with domains 19-25 [59]. On the other hand, domains 4-6 have the high-affinity binding sites for the microfibrillar protein fibrillin-1 [115], and domains 17 and 18 were identified as critical for cell adhesion to this part of TE to human dermal fibroblast via integrin αvβ5 [116]. Furthermore, a significant susceptibility of TE towards enzymatic cleavage was identified at domain 6 [83] and around domain 10 [110].

Figure 5. Model of the nanostructure of full-length and assembly of Tropoelastin.

(A) Model for full-length TE showing locations of different parts. The N-terminal region (exons 2-18) has a uniform elongated rod-shaped. It is followed by a spur region, protruding from the side of the molecule, which corresponds to a hinge region containing exons 20-24. Beyond the spur, there is a bridge to the C-terminal region that occurred at around exon 25. The molecule terminates in a more compact ‘foot-like' region (exons 26-36), which includes the cell-interactive C-terminus of TE. (B) Head-to-tail model in which a molecule of TE is covalently cross-linked via domains 19 and 25 in its tail region with the domain 10 in the head part of another TE molecule. This n-mer propagation is an outcome of the head-to-tail mechanism. The figure was adapted from Baldock et al. [9].

(27)

11 In the spur and bridge zones, it has been suggested that domains 19 and 25 are readily accessible to solvent and are enriched in intermolecular cross-links [51, 58, 59]. Domains 21/23 could be involved in the elasticity [117] and probably assist the coacervation [118] and cross-links [119]. Peptides encoded by exon 20 have a high propensity to coacervate, which could play a pivotal role in the molecular assembly of natural TE [120]. In the same way, domain 24 is also substantially solvent-exposed and could be involved in the coacervation [59]. This cross-linking enrichment could be associated with the amount of VGVAPG motif [59], which could also interact with EBP [121]. However, domains 20 and 24 played a lesser role in coacervation compared to domain 26, which has a potential role in aligning TE molecules due to a possible hydrophobic interaction with domain 18 in the intact monomer [55, 83, 112, 120]. The borders of domains 25 and 26 seem to be exposed to solvent [32] and are apparently involved in maintaining the orientation of the bridge and C-terminal region in TE during the elastic fibre assembly [122]. In particular, domain 26 is susceptible to enzymatic cleavage [83, 112].

Domains in the ‘foot-like’ part have been involved in three different roles: the cell adhesive activity, matrix interaction and TE assembly and cross-linking [9, 57, 83, 123, 124]. Domains 29 to 36 are critical assembly domains that mediate the interaction of TE with microfibrils in the ECM [83]. A crucial functional element is domain 30, which interacts with microfibrils [83, 94] and its deletion prevents the assembly of full-length TE [83, 125]. Domain 36 contains the C-terminal GRKRK motif that binds the integrin ανβ3 during the elastin fibre assembly [123, 126-128]. It supports fibroblast adhesion [106, 128] of maturing elastin [71, 124, 129]. It is possible to find cross-linking intra- or inter-molecular between domains 6 and 36 [51, 59], and cysteine residues are disulphide bonded within the molecule, precluding their direct participation in intermolecular associations [130].

1.2.5 Degradation of elastic fibres

Elastic fibres seem to undergo continuous `physiologic´ catabolism induced mainly by enzymatic proteolysis [51, 131]. Protein impairment plays a crucial function in the genesis of several diseases that involve matrix remodelling [31, 51, 132-135]. Elastases have been defined as the proteases that can degrade elastin [136], releasing peptides to

(28)

12 an appreciable extent [137]. Their proteolytic action is not exclusively limited to elastin and could encompass a variety of substrates [138]. It has been suggested that in vivo, the elastases might do a proteolytic degradation of the structural glycoprotein mantle that surrounds the elastic fibres before cleaving the elastin [137]. The enzyme activity, in

vivo, is controlled by substrate availability, enzymes involved (type, amount and

affinity), levels of their inhibitors, and the presence and function of scavenger cells [139, 140]. Aspects of the elastolysis mechanism of elastases were reviewed by Hornebeck [137].

The elastase-type proteases exhibit a broad distribution in nature [141] and have variable catalytic and substrate binding sites (For review, see [139, 142]). However, they prefer to cleave peptide bonds associated with hydrophobic or aromatic amino acids [3, 141]. Elastases are produced by different cells such as pancreatic acinar cells, polymorphonuclear leukocytes (macrophages and lymphocytes), neutrophils, mesenchymal cells, platelets, and fibroblast [134, 136, 137, 143-145]. The elastases mainly belong to three families: serine proteases: pancreatic elastase II [136], cathepsin G, human leukocyte elastase; myeloblastin (proteinase 3) [143, 145-149]; matrix

metalloproteinases (MMP-2 (Gelatinase A ), -7 (Matrilysin), -9 (Gelatinase B), and -12)

[149-154], and cysteine proteases (cathepsins K, L, V and S) [131, 134, 136, 139, 155]. Although almost all elastases are well characterised for years, only the skin fibroblast-derived elastase (SFE) remained without an undoubted identification. Recent studies indicated that SFE could correspond to the metalloprotease neprilysin (NEP) [156, 157].

1.2.6 Bioactive peptides

Along with its pivotal structural function, elastin plays a major role in the induction of specific responses from cells and tissues [126, 158]. Degradation (enzymatic or chemical) of elastin, especially of its hydrophobic domains [158], could lead to the production of elastin-derived peptides (EDPs), also called `matrikines´ or `elastokines´ [159-161]. These peptides have the ability to induce various intracellular signalling events that modify cell behaviour [162-165].

Elastokines have been detected in blood circulation in some physiological and pathological conditions, having significant physiological implications in the human health [138, 166-169]. In some cases, such as in ischemia/reperfusion injury [170, 171]

(29)

13 or tissue repair [172], EDPs might contribute to protection against damage. However, in other pathologies, EDPs seem to assist to exacerbating tissue damage. For instance, they promote emphysema [173], atherosclerosis [168, 174-177], aortic abdominal aneurysms [178], calcification of vessel walls [179], age-related macular degeneration [169], hyperplastic neointimal formation [165], melanoma [180-184] and other tumours in connective tissues [160, 185]. For reviews of the contribution of EDPs to diseases, refer to [21, 161, 186, 187].

The biological effects of EDPs have largely been attributed to their interaction with the 67-kDa elastin-binding protein (EBP). This receptor is an alternatively spliced enzymatically inactive form of the β-galactosidase that complexes with the 61 KDa and 55 kDa integral membrane proteins, carboxypeptidase A and sialidase, respectively [121, 159, 165, 188]. Its receptor is coupled with some intracellular signalling pathways that mediated EDPs cellular effect [165, 189]. Additional cell-surface receptors for EDPs have been identified. For instance, integrin ανβ3 and ανβ5 [116, 123, 128], heparin and chondroitin sulfate-containing glycosaminoglycans receptors [190] and galectin-3 receptor [191]. For elastin receptor complex review see [161] and for reviews of the interaction of cells with elastin refer to [126, 159].

It has been hypothesised that GXXP motif and a glycine residue located just after it are required for the interaction of EDPs with EBP [81, 192, 193]. These sequences are essentially inactive in the intact cross-linked protein [61]. However, elastases, even belonging to the same family, could release different peptides [137], whose biological activities are diverse and cell-specific [158]. Table 1 summarises the in vitro biological activities that have been reported for some EDPs. Furthermore, in vitro, TE has shown some cellular interactions and biological effect, which are described by Mithieux et al. See review [194].

SEQUENCE BIOLOGICAL EFFECT REFERENCE

AGLVPG;

AGLVPGGPGFGPGVV Stimulate pro-MMP-1 secretion (fibroblasts) [147]

FGVG Chemotaxis (monocytes) [195]

GAIPG Chemotaxis (M27 tumour cells) [196]

GARPG; GAVPG Stimulate pro-MMP-2 secretion (fibroblasts) [152] GFGPG; GGVLPG;

GLPGVYPGGVLPGA Stimulate pro-MMP-1 secretion (fibroblasts) [147]

GFGVG Chemotaxis (fibroblasts) [197]

(30)

14

GLVPG Chemotaxis (monocytes); stimulate pro-MMP-1

secretion (fibroblasts) [147, 195]

GVAPG Chemotaxis (monocytes) [195, 199]

GVLPG; GVYPG Stimulate pro-MMP-1 secretion (fibroblasts) [147] GYGPG Stimulate pro-MMP-2 secretion (fibroblasts) [152]

LREGDPSS Chemotaxis (monocyte) [200]

PGAIPG Chemotaxis (neutrophils, M27 tumour cells); stimulate

pro-MMP-1 and pro-MMP-3 expression (fibroblasts) [81, 196] PGFGAVPGA Stimulate pro-MMP-2 secretion (fibroblasts) [152] PGFGPG Stimulate pro-MMP-1 secretion (fibroblasts) [147] PGVGVA Stimulate elastase and superoxide release; increase

[Ca2+_]_i_{(polymorphonuclear leukocytes - PMNL)}

[201] PGVYPG Stimulate pro-MMP-1 secretion (fibroblasts) [147] VAPG Chemotaxis (WM35 and HT168-M1 melanoma cells);

stimulate of MMP-2 and MMP-3 expression (WM35 and HT168-M1 melanoma cells)

[184] VGVA Increase [Ca2+_]_i_{and endothelium vasorelaxation (vein}

endothelial cells; rat aortic rings) [202] VGVAPG Chemotaxis (endothelial cells; fibroblasts;

keratinocytes; monocyte; neutrophils; A2058, WM35 and HT168-M1 melanoma cells; M27 lung carcinoma); enhancement of atherogenesis (monocytes in mice); increase [Ca2+_]_i_{(leukocytes); induce insulin resistance}

associated with tissue remodeling (mice);inhibit proliferation (keratinocytes); myofibrillogenesis (smooth muscle cells); osteogenic response and increase expression of MMP-2 (smooth muscle cells); cell proliferation and downregulated elastin

expression (fibroblasts); promote angiogenesis (chorio-allantoic membrane); promote superoxide production and elastase release in PMNL (leukocytes); stimulate differentiation melanocyte precursors (NNCmelb4 and NCCmelan5 cells); stimulate MT1-MMP and MT1-MMP-2 expression (melanoma cells); stimulate of MMP-2 and MMP-3 expression (WM35 and HT168-M1 melanoma cells); stimulate MMP-2 expression and activation (human fibrosarcoma HT-1080 cells); stimulate pro-MMP-1 and pro-MMP-3 expression (fibroblasts); upregulation Th-1 cytokine (T-cells); vasorelaxation - increase [Ca2+_]_i_(human

umbilical vein endothelial cells-HUVEC)

[81, 164, 165, 170, 176, 178-180, 183-185, 195, 197-199, 201-213]

VGVGVA Increase [Ca2+_]_i_{; promote superoxide production and}

elastase release in PMNL (leukocytes) [201]

VGVPG Chemotaxis (monocyte) [199]

VPGVG Stimulate proliferation, inhibit elastin expression

(smooth muscle cells). [214]

VVPQ Mitogenic activity (dermal fibroblasts) [215] YGARPGVGVGGIP Stimulate pro-MMP-2 secretion (fibroblasts) [152]

YGVG Chemotaxis (monocytes) [195]

YTTGKLPYGYGPGG Stimulate pro-MMP-2 secretion (fibroblasts) [152]

Table 1. In vitro biological activities reported for some EDPs.

(31)

15 1.2.7 Elastic fibres in aging and skin diseases

Aging is understood as the gradual and progressive deterioration of integrity of multiple organ systems [216]. Loss of tissue elasticity is one of the hallmarks of aging [42, 136]. It has been associated with fragmentation and thinning of elastin structures in skin and blood vessels [42, 217], leading to loss of physiological function [218-220] and limiting the human life expectancy [221].

Cutaneous aging is a complex biological process affecting various layers of the skin [222]. It is induced by both intrinsic and extrinsic factors [223, 224]. Intrinsic (chronological) aging is genetically determined and affects the skin in a manner similar to the way various internal organs probably age. It generates fine wrinkling, increased laxity and fold accentuation, atrophy of the dermis and reduction of subcutaneous adipose tissue. Extrinsic ageing (photoaging) is induced by environmental exposure, primarily to UV radiation. Clinically it is characterised by the appearance of deep wrinkles, a sallow discoloration, telangiectasia, irregular pigmentation and furrowing and loss of elasticity [8, 131, 219, 222, 225, 226] (see Figure 6).

Structural reorganisation of the dermal ECM is evident in both protected and UV-exposed old skin [4]. In sun-protected areas, the number of elastic fibres decreased [138, 220, 227]. Actinically damaged skin is characterised by an accumulation of abnormal elastotic material in the reticular dermis as result of the solar elastosis process [220,

Figure 6. Phenotype of intrinsic ageing versus photoaging.

(A.) Intrinsic ageing skin from arms, showing an atrophy of the skin with fine wrinkling (B.) Extrinsic ageing skin displays deep wrinkling and furrowing loss of resilience and skin tone (Figure and text adopted from [8]).

(32)

16 228-230]. Some biological, biochemical and molecular mechanisms of both processes are over-imposed in sun-exposed areas [220, 231-233]. They involve the upregulation of some enzymes associated with ECM degradation [136, 218, 228, 234]. For a review of the ageing mechanism see [4, 218, 235].

Regarding the role of elastin in illness, some inherited and acquired diseases have been described to alter the structure, distribution and abundance of elastic fibres. Then, elastic fibres lose functionality and increase their susceptibility to inflammatory or proteolytic damage [16]. Broad spectra of defects are related to damages to the entire fibre and the organ system affected, including from some skeletal and skin abnormalities to vascular and ocular defects [16, 236]. Although some of these syndromes do not involve elastin as the primary target, they severely affect the elastic fibre integrity [16]. For a review see [20, 69].

The genetic diseases specifically associated with mutations in elastin gene are supravalvular aortic stenosis (SVAS; OMIM #185500), autosomal dominant cutis laxa (ADCL; OMIM #123700), and Williams-Beuren Syndrome (WBS; OMIM #194050) [20, 69, 236]. In SVAS, patients present mainly translocation and a 100 kb deletion in the 5’-end and middle region of elastin gene [237-239]. This disease is characterised by morphological and functional changes in the cardiovascular system, especially in the arterial walls [236]. ADCL is related to the single nucleotide deletion in exons 30, 32 and 34 in the 3´-end of the coding region of elastin [240-242]. As a result of this frameshift, ADCL patients present reduced elastin synthesis by skin fibroblasts and deposition in the elastic fibres, fewer microfibrils and aberrant ultrastructure of dermal elastic fibres [242, 243].

WBS is a complex developmental disorder with multisystem involvement. It is a consequence of a hemizygous contiguous gene deletion, and less frequently duplication, of ≈1.5 Mb on chromosomal band 7q 11.23, which involves several genes, usually including the elastin gene [20, 238, 244, 245]. Abnormalities in connective tissues, cardiovascular and central nervous system and at the craniofacial level are characteristic in WBS patients [244, 246, 247]. In the skin, reduction of deposition of elastin in elastic fibres, a lower diameter and a less continuous appearance of them have been described [238, 248]. Skin features such as soft consistency, premature greying of

(33)

17 the hair, while wrinkles and abnormal scarring have been detected [238, 249]. For review see [236, 246].

Acquired elastic fibre disorders arise wherever the fibre structure and function are compromised during the natural progression of the disease [20, 250]. Examples of such diseases include elastosis and several cardiovascular and pulmonary diseases, where the elastin synthesis is reinitiated to repair damaged elastic fibres [251-255]. The newly synthesised TE could be modified for some factors, such as the presence of reactive oxygen and nitrogen species, which yields to aberrant assembly and dysfunctional fibres [256]. Furthermore, EDPs could maintain a positive feedback loop of inflammation and elastin degradation; missense variants in elastic fibre proteins may predispose them to such immune-mediated elastin destruction [257].

1.3 Mass spectrometry in proteomics

Mass spectrometry (MS) is an analytical technique that has been used more than one hundred years [258]. Its basic principle is to generate ions from either inorganic or organic compounds by any suitable method, to separate these ions by their mass-to-charge ratio (m/z) and abundance [259]. Since the 1980s, MS has been applied in proteomics [258]. Currently, it is mainly used in cataloguing protein expression (molecular mass, amino acid sequence, quantitative changes), defining protein interaction and identifying sites of protein modification [260-262].

1.3.1 Instrumentation

By definition, a mass spectrometer consists of an ion source, a mass analyser and a detector [5] (see Figure 7). The measurements are carried out in the gas-phase ions of the analytical compound and their separation following their m/z [263]. Detailed review about mass spectrometers instrumentation is found in [263, 264] and a deep description of fundamentals and characteristic of MS and its instrumentation is presented in [259].

(34)

18 Separation techniques are used to decrease sample complexity before mass analysis. Some of them are gel-based (SDS-PAGE `sodium dodecyl sulphate polyacrylamide gel electrophoresis´) and liquid-based (high performance liquid chromatography (HPLC), nanoHPLC, UltraHPLC) chromatography. As they reduce the number of ions to be analysed at the same time, they help to decrease the ion suppression effect. Furthermore, chromatographic separation allows those peptides with the same m/z to elute at different retention times, and then, more ions can be identified by the mass spectrometer [265]. HPLC devices are used as sample introduction instruments in some mass spectrometers [264].

In the ion source, the sample is volatilised and ionised to generate gas-phase ions from the molecules [263, 266]. The goal of the ion sources is to generate stable molecular ions from large, non-volatile and thermally labile compounds, such as proteins and peptides, without extensive degradation [262, 263]. The two most popular ions sources techniques in proteomics are matrix-assisted laser desorption/ionisation (MALDI) and electrospray ionisation (ESI) [263] that are called soft methods due to the fact that they generate intact molecular ions from large molecules [267, 268].

Figure 7. Principal components of a mass spectrometer.

The left panel depicts the ionisation and sample introduction in (A.) matrix-assisted laser desorption/ionisation (MALDI) and (B.) electrospray ionisation (ESI). The different instrumental configurations used during this thesis (A1,2; B1,2) are shown with their typical ion source: (A1) reflector time of flight (TOF) instrument, (A2) TOF-TOF instrument, (B1) the quadrupole-TOF instrument, and (B2) the (three-dimensional) ion trap. Figure and text adapted from Aebersold and Mann et al.[5] .

(35)

19 In MALDI technique the analyte is co-crystallised with an inert matrix, such as sinapinic or α-cyano-hydroxycinnamic acid, which can absorb UV light [268]. When the dried solid mixture of matrix and analyte is irradiated with a pulsed laser, the matrix absorbs the energy and transfers it to the acidified analyte. The laser heating causes desorption of matrix and finally, singly-protonated analyte ions [M+H]+_{are released into}

the gas phase [269]. In contrast, in the ESI technique, a solution of the analyte is passed through a fine needle that has a potential difference relative to a counter electrode. The high potential induces a spray of charged droplets, which are desolvated and then, the positive or negative multiply charged ions (depending on the polarity of the applied voltage) are released [267, 270].

In the mass analyser, electric or magnetic fields are applied and change the spatial trajectories, velocity or directions of the ions, which are resolved according to their m/z [266]. The currently available analysers can be classified into two categories: (i) beam

analysers, such as quadrupole and time-of-flight analysers, in which the ions come from

the ion source in a beam and pass through the analysing field to the detector and (ii)

trapping analysers, such as ion trap, capture the ions in the analyser field where ions are

formed in the analyser itself or can be injected from an external ion source [263].

In time-of-flight (TOF) analysers, the ions are accelerated through a fixed potential into the TOF drift tube. As the velocity of the ions is inversely proportional to their masses, they are separated according to their m/z value during the movement at a constant speed [263, 264]. On the other hand, the quadrupole (Q) analyser contains four hyperbolically or cylindrically shaped rod electrodes extending in the z-direction. They are mounted in a square configuration (xy-plane). When a periodic voltage (composed of a DC voltage and a radio-frequency (RF) voltage) is applied to the rods, the overall ion motion can result in a stable trajectory causing ions of a certain m/z value (range) to pass the quadrupole [259]. Ions of different m/z can be sequentially allowed to reach the detector by increasing the magnitude of the RF and DC voltages [263].

The ion trap (IT) analysers have the electric field in all three dimensions (x, y and z), which can result in ions being trapped in the field [5, 263]. The ions are ejected from the trap successively in a mass-selective manner by increasing the RF voltage that is applied to the device [263, 264]. There are three ion traps available: three-dimensional (QIT), linear (LIT), and ion cyclotron resonance (ICR) ITs. The QIT is formed by three

(36)

20 electrodes, one ring-shaped, named annular or ring electrode, and two end cap electrodes at both sides of the ring. Inside the small cavity formed by these electrodes, the ion trapping and analysing process takes place [264]. In LIT, a trapping potential is generated by placing electrodes of slightly higher potential adjacent to the front and rear end of the multipole [263]. In ion cyclotron resonance (ICR), ions are trapped in a cell composed of four electrodes situated in a strong magnetic field, and they oscillate with a frequency (cyclotron frequency) inversely related to their m/z value [264].

The detector determines ion abundance for each corresponding ion resolved by the mass analyser according to their m/z value and generates a mass spectrum [266]. Detectors are associated with the kind of analyser used in the equipment. Some detectors used in modern mass spectrometer are the secondary electron multipliers, cryogenic detectors and the image current detection. Details about their operating mode can be found in [259].

Tandem mass spectrometry (MS/MS)

Tandem mass spectrometry (MS/MS) is the experiment used to establish the identity of a compound. MS/MS involves two stages of MS. In the first stage, ions of a desired m/z are isolated from the rest of the ions emanating from the ion source. These isolated ions (termed parent ions or precursor ions) are then induced to undergo a chemical reaction to increase their internal energy, leading to their dissociation before analysis by a second MS stage. The dissociation method almost universally used is collision-induced dissociation (CID), in which the parent ion collides with a neutral target (collision) gas and some of the kinetic energy of the parent ion can be converted to internal energy, then the dissociation happens. The instruments used for MS/MS analysis are classified in tandem-in-space or tandem-in-time. Tandem-in-space instruments require distinct analysers for each stage of MS/MS such as beam-type analysers. Trapping instruments are typically tandem-in-time, in which various stages of MS/MS are performed in the same analyser, but separated in time [263]. Mass spectrometric data acquisition is done in a data-dependent manner in which information from a current mass spectrometric scan determines the parameters of subsequent scans. In most cases, full scan produces